Researchers have introduced CoveR, a novel dense retrieval method designed to improve coverage in long-form Retrieval-Augmented Generation (RAG) systems. CoveR is trained using coverage-based contrastive and distillation objectives to better capture diverse information needs. The method was developed alongside the SCOPE dataset, which includes 90K training pairs derived from research questions and augmented with sub-question answerability signals generated by LLMs. Experiments demonstrate that CoveR boosts nugget coverage by 10% compared to existing dense retrieval baselines while maintaining relevance. AI
IMPACT Enhances information retrieval for long-form RAG, potentially improving the comprehensiveness of AI-generated content.
RANK_REASON The cluster contains an academic paper detailing a new method and dataset for information retrieval. [lever_c_demoted from research: ic=1 ai=1.0]
Read on arXiv cs.IR (Information Retrieval) →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
